wjs-mining-articles
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local utility scripts to process data and integrate with publishing workflows. It runs
scripts/parse-srt.shto clean and segment subtitle files and invokes several scripts located in the~/.claude/skills/wjs-publishing-wechat/scripts/directory (such aspangu.pyandupload-draft.sh) to format text and interface with WeChat. These operations are core to the skill's functionality. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from external SRT subtitle files provided by the user.
- Ingestion points: The file content is read by the
scripts/parse-srt.shscript and subsequently processed by the agent. - Boundary markers: While the skill employs a specific persona and formatting constraints, it does not explicitly use delimiters to isolate untrusted subtitle content from its own instructions.
- Capability inventory: The skill has the ability to write to the local filesystem and execute scripts that interact with external publishing platforms (WeChat and X/Twitter).
- Sanitization: The provided Bash and Python scripts perform structural parsing of SRT timecodes but do not filter the actual spoken text for potential malicious prompt fragments.
Audit Metadata